mcp-dns-insights
Captures DNS queries leaving your Mac and serves them to Claude via MCP for traffic pattern analysis, answering questions about top domains, categories, and temporal patterns.
README
mcp-dns-insights
Capture the DNS queries leaving your Mac and serve them to Claude over the Model Context Protocol so Claude can analyse your traffic patterns and hand back a structured JSON overview.
It answers questions like:
- Which domains does my machine talk to the most?
- How much of my traffic is advertising, analytics, or telemetry?
- When am I busiest, and were there any unusual bursts of activity?
How it works
┌────────────┐ ┌──────────────┐ ┌───────────────┐ ┌────────┐
│ tcpdump │ --> │ capture │ --> │ SQLite db │ <-- │ MCP │ <── Claude
│ (port 53) │ │ daemon │ │ (dns.db) │ │ server │
└────────────┘ └──────────────┘ └───────────────┘ └────────┘
- A capture daemon runs
tcpdump, parses each DNS question off the wire, and appends it to a local SQLite database. - An MCP server exposes read-only tools over stdio. Claude calls
analyze_patterns(and friends) and receives JSON it can summarise.
The two halves are decoupled through the database: capture is a single writer, the server is a read-only consumer.
Install
Requires Python 3.10+ and the tcpdump that ships with macOS.
git clone https://github.com/luke-nielsen/mcp-dns-insights.git
cd mcp-dns-insights
python3 -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
Usage
1. Capture DNS traffic
Packet capture needs root, so run the daemon under sudo:
sudo -E mcp-dns-insights capture
Leave it running in the background (or install it as a launchd job — see below).
Use -i en0 to pin a specific interface; the default any captures all of them.
2. Inspect from the terminal
mcp-dns-insights stats --hours 24 # human-readable summary
mcp-dns-insights stats --hours 24 --json
mcp-dns-insights info # database location & span
3. Connect it to Claude
The server speaks MCP over stdio. Register it with Claude Code:
claude mcp add dns-insights -- mcp-dns-insights serve
or add it to your MCP host config manually:
{
"mcpServers": {
"dns-insights": {
"command": "mcp-dns-insights",
"args": ["serve"]
}
}
}
Then ask Claude: "Analyse my DNS traffic from the last 24 hours and tell me what's noteworthy."
MCP tools
| Tool | Purpose |
|---|---|
analyze_patterns |
Full JSON overview: top domains, categories, temporal, notes. |
recent_queries |
The most recent queries as a JSON list. |
search_domain |
Find queries whose hostname matches a substring. |
database_info |
Database path, size, and time span. |
Shape of analyze_patterns
{
"summary": { "total_queries": 4213, "unique_hostnames": 612, "unique_domains": 241, "window": { ... } },
"top_hostnames": [ { "name": "api.github.com", "count": 88 }, ... ],
"top_domains": [ { "name": "github.com", "count": 140 }, ... ],
"query_types": { "A": 3100, "AAAA": 980, "HTTPS": 133 },
"clients": [ { "name": "192.168.1.10", "count": 4100 } ],
"resolvers": [ { "name": "1.1.1.1", "count": 4213 } ],
"categories": { "analytics": 210, "advertising": 95, "cdn": 540 },
"flagged_domains": [ { "domain": "doubleclick.net", "category": "advertising", "count": 60 } ],
"temporal": { "by_hour": { ... }, "busiest_hour": { ... }, "peak_queries_per_minute": 47 },
"observations": [ "github.com accounts for 140 queries (3% of the total).", ... ]
}
Run as a launchd job (optional)
Create ~/Library/LaunchAgents/us.radiusgroup.mcp-dns-insights.plist (capture
needs root, so a LaunchDaemon in /Library/LaunchDaemons is the better fit):
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>Label</key><string>us.radiusgroup.mcp-dns-insights</string>
<key>ProgramArguments</key>
<array>
<string>/path/to/.venv/bin/mcp-dns-insights</string>
<string>capture</string>
</array>
<key>RunAtLoad</key><true/>
<key>KeepAlive</key><true/>
</dict>
</plist>
sudo cp us.radiusgroup.mcp-dns-insights.plist /Library/LaunchDaemons/
sudo launchctl load /Library/LaunchDaemons/us.radiusgroup.mcp-dns-insights.plist
Configuration
All settings have sensible defaults and can be overridden via the environment:
| Variable | Default |
|---|---|
MCP_DNS_INSIGHTS_DB |
~/Library/Application Support/mcp-dns-insights/dns.db |
MCP_DNS_INSIGHTS_HOME |
data directory (overrides the default location) |
MCP_DNS_INSIGHTS_IFACE |
any |
MCP_DNS_INSIGHTS_FILTER |
udp port 53 or tcp port 53 |
MCP_DNS_INSIGHTS_TCPDUMP |
tcpdump |
MCP_DNS_INSIGHTS_FLUSH_INTERVAL |
2.0 (seconds) |
MCP_DNS_INSIGHTS_FLUSH_BATCH |
100 (queries) |
Privacy
The database contains a record of every hostname your Mac looked up — that is
sensitive. It is stored only on your machine and is never transmitted
anywhere except to the local MCP server you explicitly connect to Claude. The
.gitignore excludes *.db so capture data is never committed.
Note: queries answered by DNS-over-HTTPS/TLS (e.g. some browsers' built-in secure DNS) are encrypted and will not be visible to a port-53 capture.
Development
pip install -e ".[dev]"
pytest # run the test suite
ruff check . # lint
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.